Korea University's AGI Lab, directed by Professor Sungwoong Kim and Professor Sungbin Lim, for research in artificial general intelligence. We are looking for highly motivated MS and PhD students to engage in exciting research towards next-level artificial general intelligence together.

한국 고려대학교의 AGI Lab은 김성웅 교수님과 임성빈 교수님의 지도하에 인공 일반 지능 연구를 수행합니다. 차세대 인공 일반 지능을 향한 흥미진진한 연구에 참여할 의욕 넘치는 석사 및 박사 과정 학생들을 찾고 있습니다.

SGPO: Self-Generated Preference Optimization based on Self-Improver

OSPO: Object-Centric Self-Improving Preference Optimization for Text-to-Image Generation

ReactionReasoner: Towards Reasoning LLM for Chemical Reaction Prediction

Score-informed Neural Operator for Enhancing Ordering-based Causal Discovery

LM-SPT: LM-Aligned Semantic Distillation for Speech Tokenization

Slot-MLLM: Object-Centric Visual Tokenization for Multimodal LLM

Mol-LLM: Multimodal Generalist Molecular LLM with Improved Graph Utilization

ConfPO: Exploiting Policy Model Confidence for Critical Token Selection in Preference Optimization

Query-based Cross-Modal Projector Bolstering Mamba Multimodal LLM

MEVG : Multi-event Video Generation with Text-to-Video Models

Mitigating Adversarial Perturbations for Deep Reinforcement Learning via Vector Quantization

Uncertainty-Aware Rank-One MIMO Q Network Framework for Accelerated Offline Reinforcement Learning

Filling in the Gaps: LLM-Based Structured Data Generation from Semi-Structured Scientific Data

Scalable Multi-Task Transfer Learning for Molecular Property Prediction

TLCR: Token-Level Continuous Reward for Fine-grained Reinforcement Learning from Human Feedback

Efficient Latent Variable Modeling for Knowledge-Grounded Dialogue Generation

Score-based Generative Modeling through Stochastic Evolution Equations in Hilbert Spaces

Score-based Generative Models with Levy Processes

MAGVLT: Masked Generative Vision-and-Language Transformer

The Liver Tumor Segmentation Benchmark (LiTS)

LECO: Learnable Episodic Count for Task-Specific Intrinsic Reward

Selective Token Generation for Few-shot Natural Language Generation

Insights From the NeurIPS 2021 NetHack Challenge

Contrastive Regularization for Semi-Supervised Learning

Automated Learning Rate Scheduler for Large-batch Training

Hybrid Generative-Contrastive Representation Learning

Spatially Consistent Representation Learning

Visual Concept Reasoning Networks

Learning Loss for Test-Time Augmentation

AutoCLINT: The Winning Method in AutoCV Challenge 2019

torchgpipe: On-the-fly Pipeline Parallelism for Training Giant Models

Spatially Attentive Output Layer for Image Classification

Mining GOLD Samples for Conditional GANs

Fast autoaugment

Scalable Neural Architecture Search for 3D Medical Image Segmentation

Edge-Labeling Graph Neural Network for Few-shot Learning

Bayesian Model-Agnostic Meta-Learning

A Comparative Study of Modern Inference Techniques for Structured Discrete Energy Minimization Problems

Image Segmentation Using Higher-Order Correlation Clustering

A Comparative Study of Modern Inference Techniques for Discrete Energy Minimization Problems

Task-Specific Image Partitioning

Higher-Order Correlation Clustering for Image Segmentation

Large Margin Discriminative Semi-Markov Model for Phonetic Recognition

$\nu$-Structured Support Vector Machines

Wearable-Sensor Activity Analysis Using Semi-Markov Models with a Grammar

Large Margin Training of Semi-Markov Model for Phonetic Recognition

Pairwise Boosted Audio Fingerprint

Margin-Enhanced Maximum Mutual Information Estimation for Hidden Markov Models

Boosted Binary Audio Fingerprint Based on Spectral Subband Moments

The liver tumor segmentation benchmark (lits)

Edge-labeling graph neural network for few-shot learning

Bayesian model-agnostic meta-learning

A comparative study of modern inference techniques for structured discrete energy minimization problems

A comparative study of modern inference techniques for discrete energy minimization problems

Higher-order correlation clustering for image segmentation

Image segmentation usinghigher-order correlation clustering

Learning loss for test-time augmentation

Electronic device generating notification based on context data in response to speech phrase from user

Spatially consistent representation learning

Scalable neural architecture search for 3d medical image segmentation

Contrastive regularization for semi-supervised learning

torchgpipe: On-the-fly pipeline parallelism for training giant models

Pairwise boosted audio fingerprint

Wearable sensor activity analysis using semi-Markov models with a grammar

Mevg: Multi-event video generation with text-to-video models

Tlcr: Token-level continuous reward for fine-grained reinforcement learning from human feedback

Score-based generative models with Lévy processes

Taesup Kim

Ildoo Kim

Lim, Sungbin (임성빈, 林聖彬)

Chiheon Kim

Sebastian Nowozin

Jongmin Kim

Yoshua Bengio

Sungjin Ahn

Jaesik Yoon

Ousmane Amadou Dia

Pushmeet Kohli

Jörg Hendrik Kappes

Dhruv Batra

Bjoern Andres

Sungwoong Kim

In this work, we report the set-up and results of the Liver Tumor Segmentation Benchmark (LiTS), which was organized in conjunction with the IEEE International Symposium on Biomedical Imaging (ISBI) 2017 and the International Conferences on Medical Image Computing and Computer-Assisted Intervention (MICCAI) 2017 and 2018. The image dataset is diverse and contains primary and secondary tumors with varied sizes and appearances with various lesion-to-background levels (hyper-/hypo-dense), created in collaboration with seven hospitals and research institutions. Seventy-five submitted liver and liver tumor segmentation algorithms were trained on a set of 131 computed tomography (CT) volumes and were tested on 70 unseen test images acquired from different patients. We found that not a single algorithm performed best for both liver and liver tumors in the three events. The best liver segmentation …

Data augmentation is an essential technique for improving generalization ability of deep learning models. Recently, AutoAugment\cite {cubuk2018autoaugment} has been proposed as an algorithm to automatically search for augmentation policies from a dataset and has significantly enhanced performances on many image recognition tasks. However, its search method requires thousands of GPU hours even for a relatively small dataset. In this paper, we propose an algorithm called Fast AutoAugment that finds effective augmentation policies via a more efficient search strategy based on density matching. In comparison to AutoAugment, the proposed algorithm speeds up the search time by orders of magnitude while achieves comparable performances on image recognition tasks with various models and datasets including CIFAR-10, CIFAR-100, SVHN, and ImageNet. Our code is open to the public by the official GitHub\footnote {\url {https://github. com/kakaobrain/fast-autoaugment}} of Kakao Brain.

In this paper, we propose a novel edge-labeling graph neural network (EGNN), which adapts a deep neural network on the edge-labeling graph, for few-shot learning. The previous graph neural network (GNN) approaches in few-shot learning have been based on the node-labeling framework, which implicitly models the intra-cluster similarity and the inter-cluster dissimilarity. In contrast, the proposed EGNN learns to predict the edge-labels rather than the node-labels on the graph that enables the evolution of an explicit clustering by iteratively updating the edge-labels with direct exploitation of both intra-cluster similarity and the inter-cluster dissimilarity. It is also well suited for performing on various numbers of classes without retraining, and can be easily extended to perform a transductive inference. The parameters of the EGNN are learned by episodic training with an edge-labeling loss to obtain a well-generalizable model for unseen low-data problem. On both of the supervised and semi-supervised few-shot image classification tasks with two benchmark datasets, the proposed EGNN significantly improves the performances over the existing GNNs.

Due to the inherent model uncertainty, learning to infer Bayesian posterior from a few-shot dataset is an important step towards robust meta-learning. In this paper, we propose a novel Bayesian model-agnostic meta-learning method. The proposed method combines efficient gradient-based meta-learning with nonparametric variational inference in a principled probabilistic framework. Unlike previous methods, during fast adaptation, the method is capable of learning complex uncertainty structure beyond a simple Gaussian approximation, and during meta-update, a novel Bayesian mechanism prevents meta-level overfitting. Remaining a gradient-based method, it is also the first Bayesian model-agnostic meta-learning method applicable to various tasks including reinforcement learning. Experiment results show the accuracy and robustness of the proposed method in sinusoidal regression, image classification, active learning, and reinforcement learning.

Szeliski et al. published an influential study in 2006 on energy minimization methods for Markov random fields. This study provided valuable insights in choosing the best optimization technique for certain classes of problems. While these insights remain generally useful today, the phenomenal success of random field models means that the kinds of inference problems that have to be solved changed significantly. Specifically, the models today often include higher order interactions, flexible connectivity structures, large label-spaces of different cardinalities, or learned energy tables. To reflect these changes, we provide a modernized and enlarged study. We present an empirical comparison of more than 27 state-of-the-art optimization techniques on a corpus of 2453 energy minimization instances from diverse applications in computer vision. To ensure reproducibility, we evaluate all methods in the OpenGM 2 …

Seven years ago, Szeliski et al. published an influential study on energy minimization methods for Markov random fields (MRF). This study provided valuable insights in choosing the best optimization technique for certain classes of problems. While these insights remain generally useful today, the phenominal success of random field models means that the kinds of inference problems we solve have changed significantly. Specifically, the models today often include higher order interactions, flexible connectivity structures, large label-spaces of different cardinalities, or learned energy tables. To reflect these changes, we provide a modernized and enlarged study. We present an empirical comparison of 24 state-of-art techniques on a corpus of 2,300 energy minimization instances from 20 diverse computer vision applications. To ensure reproducibility, we evaluate all methods in the OpenGM2 framework and report extensive results regarding runtime and solution quality. Key insights from our study agree with the results of Szeliski et al. for the types of models they studied. However, on new and challenging types of models our findings disagree and suggest that polyhedral methods and integer programming solvers are competitive in terms of runtime and solution quality over a large range of model types.

For many of the state-of-the-art computer vision algorithms, image segmentation is an important preprocessing step. As such, several image segmentation algorithms have been proposed, however, with certain reservation due to high computational load and many hand-tuning parameters. Correlation clustering, a graph-partitioning algorithm often used in natural language processing and document clustering, has the potential to perform better than previously proposed image segmentation algorithms. We improve the basic correlation clustering formulation by taking into account higher-order cluster relationships. This improves clustering in the presence of local boundary ambiguities. We first apply the pairwise correlation clustering to image segmentation over a pairwise superpixel graph and then develop higher-order correlation clustering over a hypergraph that considers higher-order relations among superpixels. Fast inference is possible by linear programming relaxation, and also effective parameter learning framework by structured support vector machine is possible. Experimental results on various datasets show that the proposed higher-order correlation clustering outperforms other state-of-the-art image segmentation algorithms.

Learning to infer Bayesian posterior from a few-shot dataset is an important step towards robust meta-learning due to the model uncertainty inherent in the problem. In this paper, we propose a novel Bayesian model-agnostic meta-learning method. The proposed method combines scalable gradient-based meta-learning with nonparametric variational inference in a principled probabilistic framework. During fast adaptation, the method is capable of learning complex uncertainty structure beyond a point estimate or a simple Gaussian approximation. In addition, a robust Bayesian meta-update mechanism with a new meta-loss prevents overfitting during meta-update. Remaining an efficient gradient-based meta-learner, the method is also model-agnostic and simple to implement. Experiment results show the accuracy and robustness of the proposed method in various tasks: sinusoidal regression, image classification, active learning, and reinforcement learning.

In this paper, a hypergraph-based image segmentation framework is formulated in a supervised manner for many high-level computer vision tasks. To consider short- and long-range dependency among various regions of an image and also to incorporate wider selection of features, a higher-order correlation clustering (HO-CC) is incorporated in the framework. Correlation clustering (CC), which is a graph-partitioning algorithm, was recently shown to be effective in a number of applications such as natural language processing, document clustering, and image segmentation. It derives its partitioning result from a pairwise graph by optimizing a global objective function such that it simultaneously maximizes both intra-cluster similarity and inter-cluster dissimilarity. In the HO-CC, the pairwise graph which is used in the CC is generalized to a hypergraph which can alleviate local boundary ambiguities that can occur in …

Data augmentation has been actively studied for robust neural networks. Most of the recent data augmentation methods focus on augmenting datasets during the training phase. At the testing phase, simple transformations are still widely used for test-time augmentation. This paper proposes a novel instance-level test-time augmentation that efficiently selects suitable transformations for a test input. Our proposed method involves an auxiliary module to predict the loss of each possible transformation given the input. Then, the transformations having lower predicted losses are applied to the input. The network obtains the results by averaging the prediction results of augmented inputs. Experimental results on several image classification benchmarks show that the proposed instance-aware test-time augmentation improves the model’s robustness against various corruptions.

(57) ABSTRACT A method for generating a notification by an electronic device to alert a user of the electronic device is disclosed. In this method, a speech phrase may be received. Then, the received speech phrase may be recognized, by a processor, as a command to generate the notification. In addition, one or more context data of the electronic device may be detected by at least one sensor. It may be determined whether the notification is to be generated at least based on the context data. The notification may be generated, by the processor, based on the context data and the command to generate the notification.

Self-supervised learning has been widely used to obtain transferrable representations from unlabeled images. Especially, recent contrastive learning methods have shown impressive performances on downstream image classification tasks. While these contrastive methods mainly focus on generating invariant global representations at the image-level under semantic-preserving transformations, they are prone to overlook spatial consistency of local representations and therefore have a limitation in pretraining for localization tasks such as object detection and instance segmentation. Moreover, aggressively cropped views used in existing contrastive methods can minimize representation distances between the semantically different regions of a single image. In this paper, we propose a spatially consistent representation learning algorithm (SCRL) for multi-object and location-specific tasks. In particular, we devise a novel self-supervised objective that tries to produce coherent spatial representations of a randomly cropped local region according to geometric translations and zooming operations. On various downstream localization tasks with benchmark datasets, the proposed SCRL shows significant performance improvements over the image-level supervised pretraining as well as the state-of-the-art self-supervised learning methods.

In this paper, a neural architecture search (NAS) framework is proposed for 3D medical image segmentation, to automatically optimize a neural architecture from a large design space. Our NAS framework searches the structure of each layer including neural connectivities and operation types in both of the encoder and decoder. Since optimizing over a large discrete architecture space is difficult due to high-resolution 3D medical images, a novel stochastic sampling algorithm based on a continuous relaxation is also proposed for scalable gradient based optimization. On the 3D medical image segmentation tasks with a benchmark dataset, an automatically designed architecture by the proposed NAS framework outperforms the human-designed 3D U-Net, and moreover this optimized architecture is well suited to be transferred for different tasks.

Consistency regularization on label predictions becomes a fundamental technique in semi-supervised learning, but it still requires a large number of training iterations for high performance. In this study, we analyze that the consistency regularization restricts the propagation of labeling information due to the exclusion of samples with unconfident pseudo-labels in the model updates. Then, we propose contrastive regularization to improve both efficiency and accuracy of the consistency regularization by well-clustered features of unlabeled data. In specific, after strongly augmented samples are assigned to clusters by their pseudo-labels, our contrastive regularization updates the model so that the features with confident pseudo-labels aggregate the features in the same cluster, while pushing away features in different clusters. As a result, the information of confident pseudo-labels can be effectively propagated into more unlabeled samples during training by the well-clustered features. On benchmarks of semi-supervised learning tasks, our contrastive regularization improves the previous consistency-based methods and achieves state-of-the-art results, especially with fewer training iterations. Our method also shows robust performance on open-set semi-supervised learning where unlabeled data includes out-of-distribution samples.

We design and implement a ready-to-use library in PyTorch for performing micro-batch pipeline parallelism with checkpointing proposed by GPipe (Huang et al., 2019). In particular, we develop a set of design components to enable pipeline-parallel gradient computation in PyTorch's define-by-run and eager execution environment. We show that each component is necessary to fully benefit from pipeline parallelism in such environment, and demonstrate the efficiency of the library by applying it to various network architectures including AmoebaNet-D and U-Net. Our library is available at https://github.com/kakaobrain/torchgpipe .

A novel binary audio fingerprint obtained by filtering and then quantizing the spectral centroids is proposed. A feature selection algorithm, coined pairwise boosting (PB), is used to determine the filters and quantizers by casting the fingerprinting problem of identifying a query audio clip into a binary classification problem. The PB algorithm selects the filters and quantizers which lead to accurate classification of matching and nonmatching audio pairs: a matching pair is an audio pair that should be classified as being identical, and a nonmatching pair is a pair that should be classified as being different. By iteratively reducing the classification error of both matching and nonmatching pairs, the PB algorithm improves both the robustness and discriminating ability. In our experiments, the proposed fingerprint outperformed previously reported binary fingerprints in terms of robustness and discriminating ability. In the …

Detailed monitoring of training sessions of elite athletes is an important component of their training. In this paper we describe an application that performs a precise segmentation and labeling of swimming sessions. This allows a comprehensive breakdown of the training session, including lap times, detailed statistics of strokes, and turns. To this end we use semi-Markov models (SMM), a formalism for labeling and segmenting sequential data, trained in a max-margin setting. To reduce the computational complexity of the task and at the same time enforce sensible output, we introduce a grammar into the SMM framework. Using the trained model on test swimming sessions of different swimmers provides highly accurate segmentation as well as perfect labeling of individual segments. The results are significantly better than those achieved by discriminative hidden Markov models.

We introduce a novel diffusion-based video generation method, generating a video showing multiple events given multiple individual sentences from the user. Our method does not require a large-scale video dataset since our method uses a pre-trained diffusion-based text-to-video generative model without a fine-tuning process. Specifically, we propose a last frame-aware diffusion process to preserve visual coherence between consecutive videos where each video consists of different events by initializing the latent and simultaneously adjusting noise in the latent to enhance the motion dynamic in a generated video. Furthermore, we find that the iterative update of latent vectors by referring to all the preceding frames maintains the global appearance across the frames in a video clip. To handle dynamic text input for video generation, we utilize a novel prompt generator that transfers course text messages from the …

Reinforcement Learning from Human Feedback (RLHF) leverages human preference data to train language models to align more closely with human essence. These human preference data, however, are labeled at the sequence level, creating a mismatch between sequence-level preference labels and tokens, which are autoregressively generated from the language model. Although several recent approaches have tried to provide token-level (ie, dense) rewards for each individual token, these typically rely on predefined discrete reward values (eg, positive:+ 1, negative:-1, neutral: 0), failing to account for varying degrees of preference inherent to each token. To address this limitation, we introduce TLCR (Token-Level Continuous Reward) for RLHF, which incorporates a discriminator trained to distinguish positive and negative tokens, and the confidence of the discriminator is used to assign continuous rewards to each token considering the context. Extensive experiments show that our proposed TLCR leads to consistent performance improvements over previous sequence-level or token-level discrete rewards on open-ended generation benchmarks.

Investigating the optimal stochastic process beyond Gaussian for noise injection in a score-based generative model remains an open question. Brownian motion is a light-tailed process with continuous paths, which leads to a slow convergence rate for the Number of Function Evaluation (NFE). Recent studies have shown that diffusion models suffer from mode-collapse issues on imbalanced data. In order to overcome the limitations of Brownian motion, we introduce a novel score-based generative model referred to as Lévy-Itō Model (LIM). This model utilizes isotropic -stable Lévy processes. We first derive an exact reverse-time stochastic differential equation driven by the Lévy process and develop the corresponding fractional denoising score matching. The proposed generative model takes advantage of the heavy-tailed properties of the Lévy process. Our experimental results show LIM allows for faster and more diverse sampling while maintaining high fidelity compared to existing diffusion models across various image datasets such as CIFAR10, CelebA, and imbalanced dataset CIFAR10LT. Comparing our results to those of DDPM with 3.21 Fréchet Inception Distance (FID) and 0.6437 Recall on the CelebA dataset, we achieve 1.58 FID and 0.7006 Recall using the same architecture. LIM shows the best performance in NFE 500 with faster total wall-clock time than the baseline.

Professor

Fast AutoAugment

Uncertainty-aware learning from demonstration using mixture density networks with sampling-free variance modeling

A deep reinforcement learning approach for solving the traveling salesman problem with drone

Neural stain-style transfer learning using GAN for histopathological images

An Lq (Lp)-theory for the time fractional evolution equations with variable coefficients

Monte Carlo Tree Search in Continuous Spaces Using Voronoi Optimistic Optimization with Regret Bounds

Asymptotic behaviors of fundamental solution and its derivatives related to space-time fractional differential equations

Score-based Generative Models with Lévy Processes

Optimal Algorithms for Stochastic Multi-Armed Bandits with Heavy Tailed Rewards

Tsallis Reinforcement Learning: A Unified Framework for Maximum Entropy Reinforcement Learning

Generalized Tsallis Entropy Reinforcement Learning and Its Application to Soft Mobile Robots

A Sobolev space theory for stochastic partial differential equations with time-fractional derivatives

Task Agnostic Robust Learning on Corrupt Outputs by Correlation-Guided Mixture Density Networks

An L q ( L p )-Theory for Parabolic Pseudo-Differential Equations: Calderón-Zygmund Approach

Machine-learning-based diagnosis of thyroid fine-needle aspiration biopsy synergistically by Papanicolaou staining and refractive index distribution

Neural Bootstrapper

Parabolic Littlewood–Paley inequality for a class of time-dependent pseudo-differential operators of arbitrary order, and applications to high-order stochastic PDE

Kyungjae Lee

Sungjoon Choi

Hanbum Ko

Songhwai Oh

Hyungjoo Cho

Hyun-seok Min

Hyokun Yun

Taehyun Yoon

Rodrigo S.Hormazabal

Tomas Lozano-Perez

Leslie Kaelbling

Beomjoon Kim

SungYub Kim

Changhyun Kwon

Aigerim Bogyrbayeva

Sanghoon Hong

Sungbin Lim

In this paper, we propose an uncertainty-aware learning from demonstration method by presenting a novel uncertainty estimation method utilizing a mixture density network appropriate for modeling complex and noisy human behaviors. The proposed uncertainty acquisition can be done with a single forward path without Monte Carlo sampling and is suitable for real-time robotics applications. Then, we show that it can be decomposed into explained variance and unexplained variance where the connections between aleatoric and epistemic uncertainties are addressed. The properties of the proposed uncertainty measure are analyzed through three different synthetic examples, absence of data, heavy measurement noise, and composition of functions scenarios. We show that each case can be distinguished using the proposed uncertainty measure and presented an uncertainty-aware learning from demonstration …

Reinforcement learning has recently shown promise in learning quality solutions in many combinatorial optimization problems. In particular, the attention-based encoder-decoder models show high effectiveness on various routing problems, including the Traveling Salesman Problem (TSP). Unfortunately, they perform poorly for the TSP with Drone (TSP-D), requiring routing a heterogeneous fleet of vehicles in coordination—a truck and a drone. In TSP-D, the two vehicles are moving in tandem and may need to wait at a node for the other vehicle to join. State-less attention-based decoder fails to make such coordination between vehicles. We propose a hybrid model that uses an attention encoder and a Long Short-Term Memory (LSTM) network decoder, in which the decoder’s hidden state can represent the sequence of actions made. We empirically demonstrate that such a hybrid model improves upon a purely …

Performance of data-driven network for tumor classification varies with stain-style of histopathological images. This article proposes the stain-style transfer (SST) model based on conditional generative adversarial networks (GANs) which is to learn not only the certain color distribution but also the corresponding histopathological pattern. Our model considers feature-preserving loss in addition to well-known GAN loss. Consequently our model does not only transfers initial stain-styles to the desired one but also prevent the degradation of tumor classifier on transferred images. The model is examined using the CAMELYON16 dataset.

We introduce an L q (L p)-theory for the semilinear fractional equations of the type (0.1)∂ t α u (t, x)= a i j (t, x) u x i x j (t, x)+ f (t, x, u), t> 0, x∈ R d. Here, α∈(0, 2), p, q> 1, and∂ t α is the Caupto fractional derivative of order α. Uniqueness, existence, and L q (L p)-estimates of solutions are obtained. The leading coefficients a i j (t, x) are assumed to be piecewise continuous in t and uniformly continuous in x. In particular a i j (t, x) are allowed to be discontinuous with respect to the time variable. Our approach is based on classical tools in PDE theories such as the Marcinkiewicz interpolation theorem, the Calderon–Zygmund theorem, and perturbation arguments.

Many important applications, including robotics, data-center management, and process control, require planning action sequences in domains with continuous state and action spaces and discontinuous objective functions. Monte Carlo tree search (MCTS) is an effective strategy for planning in discrete action spaces. We provide a novel MCTS algorithm (voot) for deterministic environments with continuous action spaces, which, in turn, is based on a novel black-box function-optimization algorithm (voo) to efficiently sample actions. The voo algorithm uses Voronoi partitioning to guide sampling, and is particularly efficient in high-dimensional spaces. The voot algorithm has an instance of voo at each node in the tree. We provide regret bounds for both algorithms and demonstrate their empirical effectiveness in several high-dimensional problems including two difficult robotics planning problems.

Let be the fundamental solution to the problem In this paper we provide the asymptotic behaviors and sharp upper bounds of and its space and time fractional derivatives $$ D_{x}^{n}(-\Delta_x)^{\gamma}D_{t}^{\sigma}I_{t}^{\delta}p(t,x), \quad \forall\,\, n\in\mathbb{Z}_{+}, \,\, \gamma\in[0,\beta],\,\, \sigma, \delta \in[0,\infty), $$ where is a partial derivative of order with respect to , is a fractional Laplace operator and and are Riemann-Liouville fractional derivative and integral respectively.

Continuous-time score-based generative models consist of a pair of stochastic differential equations (SDEs)—a forward SDE that smoothly transitions data into a noise space and a reverse SDE that incrementally eliminates noise from a Gaussian prior distribution to generate data distribution samples—are intrinsically connected by the time-reversal theory on diffusion processes. In this paper, we investigate the use of stochastic evolution equations in Hilbert spaces, which expand the applicability of SDEs in two aspects: sample space and evolution operator, so they enable encompassing recent variations of diffusion models, such as generating functional data or replacing drift coefficients with image transformation. To this end, we derive a generalized time-reversal formula to build a bridge between probabilistic diffusion models and stochastic evolution equations and propose a score-based generative model called …

In this paper, we consider stochastic multi-armed bandits (MABs) with heavy-tailed rewards, whose p-th moment is bounded by a constant nu_p for 1

In this paper, we present a new class of Markov decision processes (MDPs), called Tsallis MDPs, with Tsallis entropy maximization, which generalizes existing maximum entropy reinforcement learning (RL). A Tsallis MDP provides a unified framework for the original RL problem and RL with various types of entropy, including the well-known standard Shannon-Gibbs (SG) entropy, using an additional real-valued parameter, called an entropic index. By controlling the entropic index, we can generate various types of entropy, including the SG entropy, and a different entropy results in a different class of the optimal policy in Tsallis MDPs. We also provide a full mathematical analysis of Tsallis MDPs, including the optimality condition, performance error bounds, and convergence. Our theoretical result enables us to use any positive entropic index in RL. To handle complex and large-scale problems, we propose a model-free actor-critic RL method using Tsallis entropy maximization. We evaluate the regularization effect of the Tsallis entropy with various values of entropic indices and show that the entropic index controls the exploration tendency of the proposed method. For a different type of RL problems, we find that a different value of the entropic index is desirable. The proposed method is evaluated using the MuJoCo simulator and achieves the state-of-the-art performance.

In this paper, we present a new class of entropyregularized Markov decision processes (MDPs), which will be referred to as Tsallis MDPs. that inherently generalize wellknown maximum entropy reinforcement learning (RL) by introducing an additional real-valued parameter called an entropic index. Our theoretical result enables us to derive and analyze different types of optimal policies with interesting properties relate to the stochasticity of the optimal policy by controlling the entropic index. To handle complex and model-free problems, such as learning a controller for a soft mobile robot, we propose a Tsallis actor-critic (TAC) method. We first observe that different RL problems have different desirable entropic indices where using proper entropic index results in superior performance compared to the state-of-the-art actor-critic methods. To mitigate the exhaustive search of the entropic index, we propose a quickand-dirty curriculum method of gradually increasing the entropic index which will be referred to as TAC with Curricula (TAC2). TAC2 shows comparable performance to TAC with the optimal entropic index. Finally, We apply TAC2 to learn a controller of a soft mobile robot where TAC2 outperforms existing actor-critic methods in terms of both convergence speed and utility.

In this article, we present an -theory ( ) for the semi-linear stochastic partial differential equations (SPDEs) of type \begin{equation*}\partial^{\alpha }_{t}u=L(\omega ,t,x)u+f(u)+\partial^{\beta }_{t}\sum_{k=1}^{\infty }\int^{t}_{0}(\Lambda^{k}(\omega,t,x)u+g^{k}(u))\,dw^{k}_{t},\end{equation*} where , and and denote the Caputo derivatives of order and , respectively. The processes , , are independent one-dimensional Wiener processes, is either divergence or nondivergence-type second-order operator, and are linear operators of order up to two. This class of SPDEs can be used to describe random effects on transport of particles in medium with thermal memory or particles subject to sticking and trapping. We prove uniqueness and existence results of strong solutions in appropriate Sobolev spaces, and obtain maximal -regularity of the solutions. By converting SPDEs driven by …

In this paper, we focus on weakly supervised learning with noisy training data for both classification and regression problems. We assume that the training outputs are collected from a mixture of a target and correlated noise distributions. Our proposed method simultaneously estimates the target distribution and the quality of each data which is defined as the correlation between the target and data generating distributions. The cornerstone of the proposed method is a Cholesky Block that enables modeling dependencies among mixture distributions in a differentiable manner where we maintain the distribution over the network weights. We first provide illustrative examples in both regression and classification tasks to show the effectiveness of the proposed method. Then, the proposed method is extensively evaluated in a number of experiments where we show that it constantly shows comparable or superior performances compared to existing baseline methods in the handling of noisy data.

In this paper we present a Calderón-Zygmund approach for a large class of parabolic equations with pseudo-differential operators of arbitrary order . It is assumed that ( t ) is merely measurable with respect to the time variable. The unique solvability of the equation and the L q ( R , L p )-estimate ∥ut∥Lq(R,Lp)+∥(−Δ)γ/2u∥Lq(R,Lp)+λ∥u∥Lq(R,Lp)≤N∥f∥Lq(R,Lp) are obtained for any λ > 0 and .

We developed a machine learning algorithm (MLA) that can classify human thyroid cell clusters by exploiting both Papanicolaou staining and intrinsic refractive index (RI) as correlative imaging contrasts and evaluated the effects of this combination on diagnostic performance. Thyroid fine-needle aspiration biopsy (FNAB) specimens were analyzed using correlative optical diffraction tomography, which can simultaneously measure both, the color brightfield of Papanicolaou staining and three-dimensional RI distribution. The MLA was designed to classify benign and malignant cell clusters using color images, RI images, or both. We included 1535 thyroid cell clusters (benign: malignancy = 1128:407) from 124 patients. Accuracies of MLA classifiers using color images, RI images, and both were 98.0%, 98.0%, and 100%, respectively. As information for classification, the nucleus size was mainly used in the color …

Bootstrapping has been a primary tool for ensemble and uncertainty quantification in machine learning and statistics. However, due to its nature of multiple training and resampling, bootstrapping deep neural networks is computationally burdensome; hence it has difficulties in practical application to the uncertainty estimation and related tasks. To overcome this computational bottleneck, we propose a novel approach called Neural Bootstrapper (NeuBoots), which learns to generate bootstrapped neural networks through single model training. NeuBoots injects the bootstrap weights into the high-level feature layers of the backbone network and outputs the bootstrapped predictions of the target, without additional parameters and the repetitive computations from scratch. We apply NeuBoots to various machine learning tasks related to uncertainty quantification, including prediction calibrations in image classification and semantic segmentation, active learning, and detection of out-of-distribution samples. Our empirical results show that NeuBoots outperforms other bagging based methods under a much lower computational cost without losing the validity of bootstrapping.

In this paper we prove a parabolic version of the Littlewood–Paley inequality for a class of time-dependent local and non-local operators of arbitrary order, and as an application we show that this inequality gives a fundamental estimate for an L p-theory of high-order stochastic partial differential equations.

AGI Lab @ KU

교수님 소개

김성웅

연구실 소개

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

SGPO: Self-Generated Preference Optimization based on Self-Improver

OSPO: Object-Centric Self-Improving Preference Optimization for Text-to-Image Generation

ReactionReasoner: Towards Reasoning LLM for Chemical Reaction Prediction

연구실 보유 특허

대학별 입시끈

우리 연구실 모집

교수님 소개

김성웅

연구실 키워드

알럼나이가 간 회사

연구실 지원 방법

대우 조건

연도별 논문 인용수

저널 & 학회 논문 출판 수

연구실 논문 목록

SGPO: Self-Generated Preference Optimization based on Self-Improver

OSPO: Object-Centric Self-Improving Preference Optimization for Text-to-Image Generation

ReactionReasoner: Towards Reasoning LLM for Chemical Reaction Prediction

연구실 보유 특허

최근 3년간 연구실적 높은 연구실

대학별 입시끈

우리 연구실 모집